Only 20 percent of students who enroll with the intent to complete in the MOOC courses offered by Stanford, Harvard and MIT, actually do so. A study of 247 classes with a quarter million students around the world, looked at four types of common intervention thought to improve that result, based on smaller earlier studies.

• Plan-making, where the students were asked to develop detailed participation and study plans;


• Value-relevance, where students were asked to write about the percieved value and relevance of the course to them;


• Benfits and barriers, asked students to postulate the benefits of completion and potential barriers to completion, then to develop plans to overcome those barriers if encountered;


• Social accountability, whereby the students were to have a person check in with them regularly to ask about their progress.

Bottom-line: Some of the interventions worked for some of the students in some of the courses. But overall, these interventions in the MOOCs did not improve course completion rates. Context matters. For example, plan-making initially helped engagement, but faded over one-two weeks. Value-relevance intervention helped in reducing the completion gap between students in developing countries and developed countries, in some courses, some of the time. The authors suggest that rather than asking 'what works?' the relevant question is 'what works, for whom, right here?'